/**
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.apache.mahout.classifier.bayes.mapreduce.cbayes;

import java.io.IOException;
import java.util.Map;

import com.google.common.collect.Maps;
import org.apache.hadoop.io.DefaultStringifier;
import org.apache.hadoop.io.DoubleWritable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.MapReduceBase;
import org.apache.hadoop.mapred.Mapper;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.util.GenericsUtil;
import org.apache.mahout.classifier.bayes.mapreduce.common.BayesConstants;
import org.apache.mahout.common.Parameters;
import org.apache.mahout.common.StringTuple;
import org.apache.mahout.math.function.ObjectDoubleProcedure;
import org.apache.mahout.math.map.OpenObjectDoubleHashMap;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * Mapper for Calculating the ThetaNormalizer for a label in Naive Bayes Algorithm
 * 
 */
public class CBayesThetaNormalizerMapper extends MapReduceBase implements
    Mapper<StringTuple,DoubleWritable,StringTuple,DoubleWritable> {
  
  private static final Logger log = LoggerFactory.getLogger(CBayesThetaNormalizerMapper.class);
  
  private final OpenObjectDoubleHashMap<String> labelWeightSum = new OpenObjectDoubleHashMap<String>();
  private double sigmaJSigmaK;
  private double vocabCount;
  private double alphaI = 1.0;
  
  /**
   * We need to calculate the idf of each feature in each label
   * 
   * @param key
   *          The label,feature pair (can either be the freq Count or the term Document count
   */
  @Override
  public void map(StringTuple key,
                  final DoubleWritable value,
                  final OutputCollector<StringTuple,DoubleWritable> output,
                  final Reporter reporter) throws IOException {
    
    if (key.stringAt(0).equals(BayesConstants.FEATURE_SUM)) { // if it is from
      // the Sigma_j
      // folder
      labelWeightSum.forEachPair(new ObjectDoubleProcedure<String>() {
        
        @Override
        public boolean apply(String label, double sigmaJ) {
          double weight = Math.log((value.get() + alphaI) / (sigmaJSigmaK - sigmaJ + vocabCount));
          
          reporter.setStatus("Complementary Bayes Theta Normalizer Mapper: " + label + " => " + weight);
          StringTuple normalizerTuple = new StringTuple(BayesConstants.LABEL_THETA_NORMALIZER);
          normalizerTuple.add(label);
          try {
            output.collect(normalizerTuple, new DoubleWritable(weight));
          } catch (IOException e) {
            throw new IllegalStateException(e);
          } // output Sigma_j
          return true;
        }
      });
      
    } else {
      String label = key.stringAt(1);
      
      double dIJ = value.get();
      double denominator = 0.5 * (sigmaJSigmaK / vocabCount + dIJ * this.labelWeightSum.size());
      double weight = Math.log1p(-dIJ / denominator);
      
      reporter.setStatus("Complementary Bayes Theta Normalizer Mapper: " + label + " => " + weight);
      
      StringTuple normalizerTuple = new StringTuple(BayesConstants.LABEL_THETA_NORMALIZER);
      normalizerTuple.add(label);
      
      // output -D_ij
      output.collect(normalizerTuple, new DoubleWritable(weight));
      
    }
    
  }
  
  @Override
  public void configure(JobConf job) {
    try {
      labelWeightSum.clear();
      Map<String,Double> labelWeightSumTemp = Maps.newHashMap();
      
      DefaultStringifier<Map<String,Double>> mapStringifier = new DefaultStringifier<Map<String,Double>>(job,
          GenericsUtil.getClass(labelWeightSumTemp));
      
      String labelWeightSumString = job.get("cnaivebayes.sigma_k", mapStringifier.toString(labelWeightSumTemp));
      labelWeightSumTemp = mapStringifier.fromString(labelWeightSumString);
      for (Map.Entry<String, Double> stringDoubleEntry : labelWeightSumTemp.entrySet()) {
        this.labelWeightSum.put(stringDoubleEntry.getKey(), stringDoubleEntry.getValue());
      }
      
      DefaultStringifier<Double> stringifier = new DefaultStringifier<Double>(job, GenericsUtil
          .getClass(sigmaJSigmaK));
      String sigmaJSigmaKString = job.get("cnaivebayes.sigma_jSigma_k", stringifier.toString(sigmaJSigmaK));
      sigmaJSigmaK = stringifier.fromString(sigmaJSigmaKString);
      
      String vocabCountString = job.get("cnaivebayes.vocabCount", stringifier.toString(vocabCount));
      vocabCount = stringifier.fromString(vocabCountString);
      
      Parameters params = new Parameters(job.get("bayes.parameters", ""));
      alphaI = Double.valueOf(params.get("alpha_i", "1.0"));
      
    } catch (IOException ex) {
      log.warn(ex.toString(), ex);
    }
  }
  
}